Email
Password
Remember meForgot password?
    Log in with Twitter

article imageOp-Ed: Computer passes itself off as human for the very first time

By Paul Wallis     Jun 8, 2014 in Technology
London - Russian software called Chatterbot has successfully passed itself off as a 13-year-old boy. This is the first time that a computer has convinced a required number of scientists that it is in fact human.
65 years ago British mathematician Alan Turing , a.k.a. the father of modern computer science, asked the question “Can machines think?” The Turing Test, which requires a computer to act and respond like a human is demanding. In this case, it was a particularly difficult type of human to imitate, too.
NBC News:
"Eugene Goostman" is not a 13-year-old boy, but 33 percent of the people who partook in five minute keyboard conversations with the computer at the Royal Society in London thought it was, according to The University of Reading, which organized the test.
The Turing Test is based on “the father of modern computer science” Alan Turing’s question, “Can Machines Think?”
If a computer is mistaken for a human by more than 30 percent of judges, it passes the test, but no computer has accomplished the feat — until now.
Think for a minute about what is required to successfully imitate a child. It’s harder than that looks. Human ears are naturally attuned to the sound of children’s voices. There is, in short, an instinctual reaction to any child. To successfully imitate a child, therefore, there is no minor achievement.
Imitating a childish personality is also a pretty thankless task. Children are not usually mistaken for computers, for some reason, and each child has its own idiosyncrasies.
The instant reactions to this major achievement were of course all over the place. One expert said that this would be a major asset for battling cybercrime, while the default response of humans at risk from technology was also evident.
Meanwhile, developers Eugene Demchenko and Vladimir Veselov are working on a very interesting dynamic which they call “conversation logic”. This dynamic, interestingly, also just happens to be the basis of human communication. Conversations are not static things. They move around, they relate to different levels of knowledge, different levels of interest, and different abilities to express thoughts and ideas.
In practice, there is one other aspect of Chatterbot which has been apparently ignored – Computers that can respond like humans can also respond to humans much more efficiently. Imagine a computer which can respond to stress, for example. A computer which recognizes anxiety attacks, depression, and others of life’s little collection of personalized sledgehammers.
The fact is that in the last 20 years computer responses have become very much part of life. Gaming, particularly, has been driving the development of artificial intelligence far beyond the original idea of artificial intelligence. People don’t mind using voice-activated software to get things done. If that software is more expressive, more responsive, and more efficient, nobody is going to mind at all.
A computer that can crack jokes, (its own jokes), hold a responsive, interesting conversation, and make sense to a human that wants to talk about a particular subject when all the other humans have run away from information overload makes sense.
Imagine a computer that can brainstorm, argue logic, and open up new areas of discussion. More importantly, imagine a computer that can keep up with streams of consciousness, which is something that’s very difficult for humans to do. Humans often can’t even keep up with their own streams of consciousness or retain them, and need to refer back to them on a regular basis. A Boswell on tap, in fact, to the Nth power.
That would be a huge breakthrough, and invaluable to just about everybody. It’s a matter of opinion how much good thinking simply passes by and disappears into the mental archives of thinkers, rarely to be seen again. It could reduce a lot of mental waste, and retain information which might otherwise be lost.
Other humans, usually, can’t do that. Shared patchy memories of a conversation or an idea talkathon often have to be glued back together. Most people need a good, responsive listener some time or other.
I’d like to add something to Turing’s original question – “Can machines be friends?” If so, humanity has created a major asset.
This opinion article was written by an independent writer. The opinions and views expressed herein are those of the author and are not necessarily intended to reflect those of DigitalJournal.com
More about Royal society, university of reading, Alan turing, can machines think, Chatterbot
More news from
Latest News
Top News